BANG - v2 (EN)
This model is a fine-tuned version of openai/whisper-large-v3 on the Radio-Modified Common Voice 11.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.2650
- Wer: 20.5610
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
0.7709 | 0.25 | 1000 | 0.6383 | 35.6607 |
0.4424 | 1.2443 | 2000 | 0.4248 | 26.8037 |
0.2823 | 2.2385 | 3000 | 0.3117 | 22.4425 |
0.2429 | 3.2328 | 4000 | 0.2650 | 20.5610 |
Framework versions
- Transformers 4.41.1
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1
- Downloads last month
- 2
Finetuned from
Dataset used to train michaelszhu/whisper-small-fintuned-radio-ASR
Evaluation results
- Wer on Radio-Modified Common Voice 11.0test set self-reported20.561